The Benefits of Hybrid Caching in Gauss–Poisson D2D Networks

Device caching has recently been proposed as an efficient way to offload traffic from congested cellular networks. However, previous works usually ignore the fact that, in practice, the user device may not be willing to help others due to the limited battery capacity. In this paper, we introduce cooperation among the device-to-device (D2D) transmitters and propose two novel hybrid caching strategies-single-point caching combined with two-point cooperative caching with joint transmission (SPC-CCJT) or multi-stream transmission (SPC-CCMT)-aiming at saving the energy cost of content deliverers. Using tools from stochastic geometry, we propose an analytical framework of the hybrid caching strategies by modeling the locations of the D2D transmitters as a Gauss-Poisson process (GPP) to accurately capture the clustering and cooperative behaviors. First, we consider a probabilistic caching placement and optimize the caching distribution to maximize the cache hit probability. Second, to compare the performance between different content delivery strategies, we derive the success probability and per-user capacity for SPC, CCJT, and CCMT, respectively. These results are then applied to evaluate the offloading gain and the distribution of the content retrieval delay for SPC-CCJT and SPC-CCMT in the GPP-based D2D networks. It turns out that significant offloading gain and delay improvement can be achieved by hybrid caching with cooperation while the energy cost of each cooperator is kept low.